Development of scalable manufacturing routes to AZD1981. Application of the Semmler-Wolff aromatisation for synthesis of the indole-4-amide Core

Article abstract of DOI:10.1021/op300173z

A safe and efficient synthesis of AZD1981 is described in which the indole 4-amide core is formed by a Semmler-Wolff aromatisation of a cyclohexenone oxime fused to a pyrrole ring. The substrate was obtained via Paal-Knorr pyrrole synthesis, followed by i

Full text of DOI:10.1021/op300173z

Article
pubs.acs.org/OPRD
Development of Scalable Manufacturing Routes to AZD1981.
Application of the Semmler−Wolff Aromatisation for Synthesis of the
Indole-4-amide Core
Manjunatha Sulur,§ Parhalad Sharma,*,§ Ravi Ramakrishnan,§ Ravi Naidu,§ Eric Merifield,†
Duncan M. Gill,†,⊥ Adrian M. Clarke,† Colin Thomson,† Michael Butters,‡ Sreekanth Bachu,§
Colin H. Benison,‡ Nagaraju Dokka,§ Emily R. Fong,† David R. J. Hose,‡ Gareth P. Howell,†
Siobhan E. Mobberley,† Simon C. Morton,† Alexander K. Mullen,‡ Jayan Rapai,§ and Bharathi Tejas§
†Pharmaceutical Development, AstraZeneca R & D Charnwood, Bakewell Road, Loughborough, Leicestershire, LE11 5RH, United
Kingdom
‡Pharmaceutical Development, AstraZeneca, Charter Way, Macclesfield, SK10 2NA, United Kingdom
§Pharmaceutical Development, AstraZeneca India Pvt. Ltd., Hebbal, Off Bellary Road, Bangalore 560024, India
S
* Supporting Information
ABSTRACT: A safe and efficient synthesis of AZD1981 is described in which the indole 4-amide core is formed by a Semmler−
Wolff aromatisation of a cyclohexenone oxime fused to a pyrrole ring. The substrate was obtained via Paal−Knorr pyrrole
synthesis, followed by incorporation of the key 3-arylthio substituent by reaction with 4-chlorophenylsulfenyl chloride. In this
manner, the 1,2,3,4-substitution pattern of the AZD1981 core was regiospecifically established in a concise and efficient
telescoped sequence. Accordingly, AZD1981 was obtained in 40% overall yield in six chemical steps, with two isolated crystalline
intermediates.
could be a potential safety issue. In addition to this, there was
inconsistency in the yield of 2 due to poor control over the
Makosza chemistry. This prompted a search for alternative
means of manufacturing 1 that did not proceed through 2. The
invention and development of a next generation synthetic route
to AZD1981 is the subject of this paper.
INTRODUCTION
■
AZD1981 (1) is being developed in the Respiratory and
Inflammation portfolio within AstraZeneca.1 The evolution of
the first generation supply route, which relied upon a sequential
elaboration of 2-methyl-4-nitro-1H-indole (2), itself was
prepared through Makosza reaction from 3-nitroaniline2
(Scheme 1).
Having critically evaluated a number of established routes to
substituted indoles that relied on annelating a benzenoid ring,
an alternative construction which established the five-
membered ring prior to elaboration of the benzenoid ring
appeared to offer a superior opportunity to avoid the
intermediacy of 2. The Semmler−Wolff aromatisation is
particularly appealing in this connection.3This reaction provides
a method for conversion of cyclohexenone oximes into
aromatic anilines or anilides by dehydration under acidic
conditions.4 Whereas the majority of published examples relate
to the synthesis of isolated benzenoid rings, a number of fused
heterocyclic systems have also been prepared by this method,
including an amino carbazole,5 an oxindole-4-amide,6 and a 4-
amino benzoxazole.7To the best of our knowledge, this reaction
has not been applied to the synthesis of an indole, but were it
to prove feasible, it would enable the direct preparation of the
4-amidoindole core of AZD1981, as illustrated in Scheme 2.
To minimise the overall number of steps, our goal was to use
a substrate which allowed early introduction of the indole N-
substituent from a glycine derivative and for concomitant
acetylation of the amino group during the Semmler−Wolff
Scheme 1. AZD1981 Early Development
Although this route proved robust and scalable in early
development, we had concerns over the long-term economic
viability of the route and certain aspects of process safety
regarding specifically the preparation and thermal stability of 2
which failed the Koenen tube test (2 mm diameter).
Moreover, Makosza chemistry for the synthesis of 2 is carried
out above the flash point of acetone in the presence of air which
Received: June 29, 2012
Published: October 18, 2012
© 2012 American Chemical Society
1746
dx.doi.org/10.1021/op300173z | Org. Process Res. Dev. 2012, 16, 1746−1753

Organic Process Research & Development
Article
crystallised. Ethyl acetate and isopropyl acetate proved
acceptable for this purpose; however, seeding was found
necessary to generate the product with an adequate physical
form. Demonstration of the ethyl acetate process by scale up to
500 g scale (input dione) afforded 9 in 62% yield. However,
MTBE ultimately proved a better choice of cosolvent since the
product could be crystallised from this solvent without seeding.
The process defined for scale up used a small excess of
chloroacetone and required a reaction time of 20 h at 30−35
°C to achieve less than 3% (area by HPLC) of 7 remaining.
Following addition of MTBE, sodium acetate, and glycine ethyl
ester, pyrrole formation took place over 6−8 h at 50−55 °C. A
sodium bicarbonate wash was introduced into the workup to
remove a small amount of the de-esterified product, which was
routinely observed as a low level impurity. On a 130 kg input
scale, by this process, 9 was produced in overall yield of 62−
Scheme 2. Semmler−Wolff Synthesis of Indole-4-amides
reaction itself. One development aspect of this alternative route
to AZD1981 was defining the order in which the trans-
formations would be carried out, particularly the point at which
the Semmler−Wolff aromatisation would be implemented.
Consequently, a choice between using either substrate 3 or 4
was a key consideration in establishing this new approach.
Whereas conditions for sulfenylation of indole 5 had already
been established, the possibility of side reactions of 3 with
electrophilic acylating agents was apparent. To mitigate this
risk, compound 4 was targeted as a key substrate.
b
67% with very high purity.
Semmler−Wolff Aromatisation of 3. Conditions tradi-
tionally used for the Semmler−Wolff aromatisation involve
heating the oxime substrate and acetic anhydride in acetic acid
whilst passing gaseous hydrogen chloride through the mixture,
and their use for the preparation of acetanilides is well
precedented.9 In combination with acetic anhydride, other acids
such as phosphoric10 and methane sulfonic11 acids promote the
reaction under milder conditions. Whilst acetic anhydride
alone,12 or in the presence of acetic acid,13 has been used,
reported reaction times are lengthy. Alternatively, acetyl
chloride with acetic anhydride in the presence of pyridine,14
ketene,15 or 1-ethoxyvinyl acetate16 promoted the Semmler−
Wolff aromatisation under milder conditions. The proposed
mechanism of Semmler−Wolff reaction encompasses several
individual steps, commencing with O-acetylation of the oxime.
Subsequent aromatisation occurs by loss of acetic acid together
with N-acetylation.17
RESULTS AND DISCUSSION
■
Synthesis of the Pyrrolocyclohexenone Substrate. A
straightforward route to the required pyrrolocyclohexenone 9
was devised utilising a Paal−Knorr pyrrole synthesis on trione
8, itself made in a straightforward manner by alkylation of
cyclohexane-1,3-dione 7 using chloroacetone, as illustrated in
Scheme 3.8 Initial studies into the synthesis of 8 using sodium
Scheme 3. Synthesis of the Pyrrolocyclohexanone 9
Development of Semmler−Wolff transformation of 3 started
with a one-pot “all-in” process (Scheme 4) in which the
or potassium hydroxides as base in aqueous ethanol confirmed
the extensive reaction time reported. Under such conditions, at
ambient temperature, around 80% conversion to 8 was
achieved after 20 h. It was possible to accelerate the reaction
by switching to potassium tert-butoxide in neat ethanol, but
these conditions led to formation of unacceptably high levels of
furan 9a.
Scheme 4. Semmler−Wolff Reaction of 3
We studied the reaction using in situ mid-IR monitoring
(employing aqueous ethanol as reaction solvent and potassium
hydroxide as base). This demonstrated that the rate of reaction
was essentially independent of solvent composition over quite a
wide range at 20 °C. Further formation of 9 by reaction of 8
with glycine ethyl ester hydrochloride in the presence of
sodium acetate at elevated temperature was found to be highly
tolerant of water and potassium chloride. Accordingly, neither
isolation of 8 nor drying solutions of this intermediate was
imperative for a successful pyrrole synthesis. This encouraged
us to look at water alone as reaction solvent for the preparation
of 8. Under these conditions, an acceptable 85 area %
conversion was achieved after an overnight stir at 20 °C.
reaction was allowed to proceed to its full extent without
isolation of any intermediate. The required oxime 3 was
synthesized by refluxing ketone with hydroxyl amine hydro-
chloride and sodium acetate in ethanol with overall yield of
54%. To begin with, 3 was subjected to many typical reported
conditions (Table 1) for the Semmler−Wolff aromatisation,
aiming for in situ N-acetylation to get 5.
As shown in Table 1, palladium/carbon as a catalyst gave
very little conversion for the desired product along with many
impurities (entry 1). Other conditions mainly resulted in imide
product instead of amide (entry 3−7). With acetic anhydride
alone as reagent, the reaction was found to require quite harsh
conditions, e.g., temperatures of 140−160 °C, to achieve some
conversion (entry 2), although shorter reaction times were
achievable using microwave heating at similar temperatures
(entry 7). In principle, the requisite amide 5 could be
a
However, the product was observed to separate out as oil and
could not be recovered completely in this form or by extraction
into a water-immiscible solvent. Therefore, the aqueous−
organic mixture was progressed directly into the Pall−Knorr
pyrrole synthesis following addition of an organic cosolvent.
This facilitated removal of the aqueous potassium chloride at
the end of the reaction whilst enabling isolation of 9 by
partition into the organic phase, from which it could be
1747
dx.doi.org/10.1021/op300173z | Org. Process Res. Dev. 2012, 16, 1746−1753

Organic Process Research & Development
Article
ethylacetate; however, the heterogeneous nature of the
resulting reaction mixture prompted replacement by sulfuryl
chloride in the same solvent. Complete conversion of 10 under
these conditions was confirmed by undertaking a study in D8-
EtOAc and analysis of the mixture directly by 1H NMR
spectroscopy, which provided a spectrum consistent with the
formation of the sulfenyl chloride. Reaction of 11 with pyrrole
9 conveniently also proceeded in ethyl acetate, providing
thioether 12 as a solution in ethyl acetate (Scheme 5). Under
such conditions, thioether 12 could be obtained in 80% yield by
solution assay on a small laboratory scale.
Table 1. Initial Screening for the Aromatising Reagents and
Catalysts
5 by
LC
13 by
LC
aromatising reagents/
catalyst
entry
1
area %
area %
remarks
a
Pd/C
19
---
Aromatisation with
many impurities
b
2
3
Ac2O
35
3
2
Low conversion
c
TFAA
53
Aromatisation but
mainly imide 13
d
4
5
6
7
Ac2O/HBr/xylene
4
4
45
45
53
54
Aromatisation with
many impurities
e
Ac2O/thiol/xylene
Major imide with
other impurities
Scheme 5. Synthesis of the Semmler−Wolff Substrate
f
Ac2O/AcOH
3
Closed vial reaction,
but mainly imide 13
g
Ac2O/microwave
24
Aromatisation with
major imide
component
8
9
xylene/S-phenyl
h
60
66
77
---
---
---
48% isolated yield
Product not isolated
45% isolated yield
thioacetate (2.0
equiv)
i
1-acetyl imidazole (2.0
equiv) and
thiophenol
j
10
Ac2O/xylene/TBAI
a
b
c
Diethylene glycol dimethylether/150−180 °C. Reflux. Neat/110
d
e
f
g
h
°C. 110 °C. 110 °C. 110 °C. 145 °C/40 min. Xylene/140 °C/40
min. 140−150 °C. 110 °C.
i
j
Isolation of intermediate 12 by crystallisation proved difficult,
and low recoveries resulted. These observations focused
development upon a telescoped process whereby a solution
of thioether 12 in ethyl acetate was progressed directly into
oxime formation. A basic aqueous workup was introduced
principally to remove the HCl produced, before transition into
the oxime formation. Oximation being water sensitive,
concentration of thioether solution in ethyl acetate by
distillation dried the solution sufficiently to enable efficient
conversion to the oxime 4. To begin with, this was achieved
quite simply by heating a solution of the substrate in ethyl
regenerated from the imide 13 under the basic conditions, but
in practice, this complicated isolation of intermediates and
impacted yield through losses to liquors. Therefore, minimising
the formation of imide 13 became an important objective of the
development studies.
An initial breakthrough arose during evaluation of alternative
acetylating agents in place of acetyl chloride or acetic
anhydride. Use of S-phenyl thioacetate in xylene (entry 8)
was found to give better conversion than acetic anhydride alone
under comparable reaction conditions and provided the starting
point for a catalyzed Semmler−Wolff aromatisation reaction.
Noting that a reaction using acyl imidazole (entry 9) was
enhanced upon addition of thiophenol, the hypothesis that the
thiol may be acting as a nucleophile, either by activating the
acetylating agent or through having an alternative role in the
reaction, prompted investigation of iodide salts as promoters
(entry 10).18
Indeed, when using acetic anhydride in xylene in the
presence of tetrabutylammonium iodide (TBAI), a good
conversion of oxime 3 was observed over a short period of
time (3−4 h) at reflux to form indole amide 5 cleanly. The
removal of TBAI had some problems, and finally the product
was isolated with 45% yield.
Though the results were encouraging, the formation of 13 in
considerable amount along with some of the unidentified
impurities raised concerns about control of product quality at
scale. This prompted examination of substrate 4 under similar
conditions.
c
acetate or ethanol with hydroxylamine hydrochloride in the
presence of sodium acetate. However batches of oxime
prepared this way tended to have low and variable assays,
presumably due to contamination with inorganic materials.
Therefore, a screen of common inorganic and organic bases was
carried out and identified tri-n-butylamine as a better
alternative, for which the soluble hydrochloride salt remained
in solution upon isolation of oxime 4, in 53 and 54% yields
from either ethyl acetate or ethanol, respectively. Upon the
basis of the results that oxime formation worked equally well in
either ethyl acetate or ethanol alone, use of a mixed solvent
system in fact provided the simplest process.
Thus, following thioether formation in ethyl acetate and
concentration, ethanol was added, followed by hydroxylamine
hydrochloride and tri-n-butylamine. Following a 4 h stir-out at
60 °C and cooling to 20 °C, 4 was collected by filtration and
obtained in around 80% yield on a small laboratory scale.
Investigation of various reaction parameters for the two-stage
telescope was undertaken by a single combined factorial
experimental design (FED) prior to the first pilot plant scale up
and covered, amongst other aspects, temperatures, volumes,
stoichiometries, addition, rate, and solvent composition for the
oxime stage (Table 2).
Synthesis of Oxime 4. Conversion of thiols or disulfides to
the corresponding sulfenyl chlorides has precedence using
reagents such as sulfuryl chloride or chlorine. Indeed, 4-
chlorophenylsulfenyl chloride 11 has been prepared by such
means from both the thiol19 and disulfide.20 Of these, the easily
handled solid disulfide 10 was selected due to ready availability
with very good purity. Initially, chlorination of this substrate
was carried out using trichloroisocyanuric acid (TCCA) in
As noted above from Table 2, the stoichiometry of sulfuryl
chloride and disulfide 10 was matched on a molar basis, as was
that for tributylamine and hydroxylamine hydrochloride. The
sulfenyl chloride formation and reaction with pyrrole 9 were
1748
dx.doi.org/10.1021/op300173z | Org. Process Res. Dev. 2012, 16, 1746−1753

Organic Process Research & Development
Article
reaction required an elevated temperature (110 °C) to proceed
at a reasonable rate. Unfortunately, the downside of using a
homogeneous source of iodide was separation from the product
at the end of the reaction, which was exacerbated by the large
amount (up to 1 mol equivalent) used (entry 1). To facilitate
removal of the additive, inorganic iodide salts (potassium
iodide, entry 2, and sodium iodide, entry 3) were assessed, and
these compared favorably during small lab-scale trials in terms
of reaction profile. Further high levels of imide 15 formation
from use of a large excess of acetic anhydride (entry 4) as both
reagent and solvent demonstrated the requirement for inert
diluents. Those which were shortlisted, having met the required
criteria, were acetic acid, anisole, mesitylene, xylene, toluene,
and 2-methyltetrahydrofuran as shown in Table 3 (entries 5−
9).
Table 2. Factorial Experiment Design (FED) Range versus
Proposed Conditions
proposed
entry
1
factors
design range
condition
sulfenyl chloride mol equiv
0.5−0.55 mol
0.53
equiv
2
3
4
temperature of ketone formation
0−40 °C
0.9−1.1 equiv
5−9 rel. vol.
0 °C
1.0 equiv
6 rel. vol.
NH2OH·HCl and Bu3N equiv
total solvent volume for oxime
formation
5
6
7
solvent composition for the oxime 33−67% v/v
33% v/v
EtOH
formation
EtOH
temperature for the oxime
formation
50−70 °C
60 °C
temperature of the oxime isolation 10−30 °C
20 °C
In acetic acid, complete consumption of acetylated oxime 14
was achieved within 18 h at 80 °C (entry 10). However, use of
a 50:50 xylene/acetic acid mixture at 95 °C (entry 11) achieved
complete reaction within 2 h with better purity when compared
with reactions in xylene alone. Further assessment of mixed
solvent systems with acetic acid led to the conclusion that
xylene was indeed the nonpolar component of choice since it
was only surpassed by mesitylene, which was less readily
available and potentially more difficult to remove from the
product due to its higher boiling point. The acetic anhydride
stoichiometry was a balance between benefit on rate of reaction
and controlling overacetylation to generate imide 15, and
consequently 4 mol equiv was used, whereby a reaction time of
4−6 h was observed at a temperature of 85 °C. Up until this
point in time, the amount of catalyst was maintained at 50 mol
%, and the main benefit of such a high level was a short time to
reach end of reaction. By implementing less of the additive (10
mol %), it appeared that substantial removal from reaction
mixtures during work up was now possible, and although the
reaction time was increased, the ultimate extent of reaction
remained essentially unchanged.
best done at low temperature (0 °C). The stoichiometry of
hydroxylamine hydrochloride was chosen to keep the excess of
this reagent to a minimum. In this manner, sulfenyl chloride 11
reacted with pyrrole 9 to afford thioether 12 cleanly in 90%
yield by solution assay on a large laboratory scale. Subsequent
conversion to oxime 4 at a reaction temperature of 60 °C
achieved complete reaction in 4 h in a solvent system
comprising a 2:1 mixture of ethyl acetate and ethanol. Oxime
4 was isolated simply by filtration at the end of reaction in 88%
yield. Upon scale up of these conditions to pilot plant scale (90
kg input of pyrrole 9), the yield was 84% with good purity.
Semmler−Wolff Aromatisation of 4. Applying here our
learnings of Semmler−Wolff aromatisation of 3, an initial
screening of reaction of 4 was done using acetic anhydride in
xylene using TBAI as promoter (Scheme 6, Table 3). The
Scheme 6. Semmler−Wolff Aromatisation
Nevertheless, having optimised the solvent composition, re-
evaluation of sodium iodide at a level of 5 mol % demonstrated
a comparable time to reach the end of reaction (entry 12). An
important observation was made during the course of this work
in response to variability that was impacting time to achieve
end of reaction and in some cases making the reactions stop
with unusually high levels of substrate remaining. This was
attributed to inadequate inertisation of the reaction, which
perhaps resulted in oxidation of the iodide catalyst to iodine to
variable extents. By inertisation of the reaction vessels prior to
charging the reagents, reproducibility was ensured.
As a result of some concerns over degradation of acetyl
oxime 14 at elevated temperature, the effect of separating out
the initial oxime acetylation step from the remainder of the
transformation was studied through investigation of a two-stage
process. To accomplish this, a solution of acetylated oxime 14
was prepared in the first reaction vessel in a mixture of xylene
and acetic acid at room temperature containing part of the
acetic anhydride charge. This was subsequently added to a hot
mixture of the remaining acetic anhydride, and the sodium
iodide in the same mixed solvent system held at the required
reaction temperature in the second vessel. This change resulted
in a faster reaction and an increased yield in solution measured
by assay at end of reaction from 83% to 93% and was therefore
adopted as the standard method by which the process was
operated.
Table 3. Series of Experiments to Compare Certain Aspects
of the Reaction Conditions
6 by LC
a
entry
1
conditions
area %
remarks
Ac2O/TBAI/110 °C
87.4
TBAI removal needed
purification using IPA
2
3
Ac2O/KI/95−100 °C
Ac2O/NaI/85−95 °C
Ac2O/
92.3
86.3
3.0
70
Isolated yield 65%
Isolated yield 65%
4
15 (imide) >80% by LC area
15 (imide) 8.9% by LC area
15 (imide) 6.9% by LC area
15 (imide) 7.6 by LC area
15 (imide) 5.3% by LC area
15 (imide) 6.9% by LC area
b
5
Ac2O/xylene
b
6
Ac2O/toluene
71
b
7
Ac2O/anisole
67
b
8
Ac2O/mesitylene
64
b
9
2-MeTHF/Ac2O
68
10
xylene:AcOH(1:1)/NaI
(10%)/80 °C
90
18 h for the reaction
completion
11
12
xylene:AcOH(1:1)/NaI
93
93
Unreacted 4 less than 1.0%
(10%)/95 °C
by LC area after 2 h
xylene:AcOH(1:1)/NaI
(5%)
86% solution yield
a
Note: entries 5−9: Ac2O(4.0 mol equiv), solvent (5 rel. vol.),
b
temperature 107 °C (oil bath). No catalyst used.
1749
dx.doi.org/10.1021/op300173z | Org. Process Res. Dev. 2012, 16, 1746−1753

Organic Process Research & Development
Article
Further development of the reaction through investigation of
process parameters and stoichiometry prior to pilot plant scale
up was done by factorial experimental design (FED),
encompassing solvent composition, amount and split of the
acetic anhydride charge over the two parts of the process, and
addition rate of the acetyl oxime to the hot reaction mixture
(Table 4). This confirmed that an acetic acid rich solvent
The downstream chemistry of 6 to get AZD1981 (1) by de-
estrification and further purification by crystallisation remains
the same as in the case of the early developmental route.4
However, an attempt to replace IPA with EtOH at the
hydrolysis stage was successful and afforded AZD1981 (1) in
94% yield with a purity of 99.5 area % on a 100 kg input scale.
The overall yield of 1 from the Semmler−Wolff aromatisation
route is about 37−40% against 18−35% of the first generation
supply route.
Table 4. Study on Process Parameters and Stoichiometry by
Factorial Experimental Design (FED)
CONCLUSIONS
■
entry
1
parameters
reaction
temperature
range studied
90−115 °C
proposed conditions
A highly efficient, robust, and safe synthesis of AZD1981 from
cheap, readily available building blocks, and which is suitable for
large-scale manufacture, has been devised (Scheme 7). The key
100 °C
2
3
mol equiv of NaI 0.025−0.1 mol equiv
0.05 mol equiv
addition rate of
acyl oxime
0.05−0.3 mL/min
addition rate: no
impact on quality
Scheme 7. Overall Synthetic Scheme of AZD1981
4
5
Ac2O mol equiv 1.2−2.8 equiv (split
2.5 mol equiv
between two vessels)
% of xylene wrt 30−90% v/v wrt AcOH
50:50
AcOH
system was preferred; however, to balance this against the
impact on yield of the isolated product, the proportion of this
was limited to 50% v/v. The split of the acetic anhydride charge
across the two reaction vessels (initially set at 3.8 mol equiv)
was not important, as long as at least 1 equiv was present in
vessel 1 to acetylate the oxime; therefore, an equal split was
chosen. The amount of acetic anhydride was subsequently
reduced to 2.5 mol equiv to better control over-reaction to
imide 15, but at the expense of a slightly lower yield in solution
(87%) and longer reaction time. The reaction temperature for
the second part of the process was increased to 100 °C, without
detriment, to ensure a good margin above the lower limit of 95
°C, below which the rate of reaction was significantly impacted.
Isolation of the product with a good recovery depended
upon effective removal of the acetic acid since high residual
levels at the point of crystallisation led to loss of the product to
liquors. Removal through washing was preferred over
distillation, and consequently aqueous sodium chloride and
dilute sodium thiosulfate washes were implemented to remove
acetic acid and iodine arising from adventitious oxidation of the
iodide salt. It was essential to carry out the workup at elevated
temperature to maintain the product in solution, and after
addition of heptane and cooling, crude amide 6 was isolated
with a good physical form. Under these conditions on a 100 g
laboratory scale, amide 6 was formed in 87% yield, measured by
solution assay, and isolated in 76% yield. Scale up of these
reaction conditions to pilot plant (73 kg input of oxime 4)
afforded the product in 76% yield.
Two effective systems for recrystallisation of amide ester
were identified from screening studies, ethanol or aqueous
acetonitrile, and the latter was chosen as a solvent of choice.
The recrystallisation was scaled up successfully, affording
purified indole amide 6 in 92% recovery.
Further development of this stage therefore focused on use
of around 10% v/v acetic acid, which offered product with
better purity and allowed a simpler work up by removal of the
first aqueous wash, in which around 4% yield was lost, and
necessitating only the dilute aqueous sodium thiosulfate wash
to remove iodine. Further optimisation of the reaction volumes
and isolation procedure arrived at the improved process which
afforded indole amide 6 in 81% yields on a 200 g scale.
step, a Semmler−Wolff aromatisation, has been found to
proceed under relatively mild reaction conditions using acetic
anhydride in a mixture of acetic acid and xylene in the presence
of a catalytic amount of sodium iodide. The solvent
composition has been optimized. The exact role of the iodide
has not yet been elucidated and is the subject of ongoing work
within our laboratories.
EXPERIMENTAL DETAILS
■
General. All reactions were performed under a nitrogen
atmosphere. Xylene refers to a commercially available mixture
of o-, m-, and p-xylenes and ethyl benzene. Heptane refers to n-
heptane (assay 99% w/w minimum). NMR spectra were
recorded on Varian Unity Inova spectrometers operating at the
specified proton frequencies on D6-DMSO solution with
tetramethylsilane (TMS) as an internal reference. HPLC and
UHPLC methods which can be used for purity and assay
determination are given in the Supporting Information. Assays
were performed against purified and fully characterised
reference materials. LC−MS spectra were obtained using an
Agilent 1100 series LC instrument and LC/MSD SL detector
with +ve atmospheric pressure electrospray ionisation (APESI).
Ethyl (2-Methyl-4-oxo-4,5,6,7-tetrahydro-1H-indol-1-
yl)acetate (9). To a 1600 L glass-lined reactor was charged
water (143 L) and cyclohexane-1,3-dione (130 kg, 1.16 kmol)
at 27 °C, and then the mixture was cooled to 12 °C with
stirring. To this, a solution of potassium hydroxide (76.6 kg,
90% w/w, 68.9 kg contained weight, 1.23 kmol) in water (117
L) was added over a period of 2 h 15 min at 11 °C (initial pH
5.0 increasing to 7.0 at the end of the addition), and then the
mixture was maintained at this temperature for a further 30
min. Chloroacetone (126.2 kg, 1.36 kmol) was added slowly to
1750
dx.doi.org/10.1021/op300173z | Org. Process Res. Dev. 2012, 16, 1746−1753

Organic Process Research & Development
Article
the reaction mixture over 4 h at 11.5 °C, after which the
temperature was allowed to rise to 32 °C and maintained for 20
h. Glycine ethyl ester hydrochloride (180.7 kg, 1.29 kmol) and
MTBE (390 L) were added, and after continuing stirring at 30
3 °C for 10 min, anhydrous sodium acetate (104.7 kg, 1.28
kmol) and water (260 L) were charged. Then the reaction
been collected. After cooling the concentrate to 20 °C,
hydroxylamine hydrochloride (26.1 kg, 376 mol) and tributyl-
amine (69.8 kg, 376 mol) were added followed by ethanol
(140.9 kg), and then the resulting mixture was heated at 60
3
°C for 3.5 h. After cooling to 20 °C, the solid was collected by
centrifugation, washed with ethyl acetate (218 kg), and then
dried under vacuum at a maximum jacket temperature of 40 °C
to provide the title compound as a white solid, 125.8 kg (84%
yield). Purity by HPLC 96.3 area %. Assay by HPLC 97% w/w.
1H NMR (400 MHz): 10.30 (1H, s), 7.21−7.25 (2H, m),
6.89−6.93 (2H, m), 4.85 (2H, s), 4.17 (2H, q, J = 7.1 Hz),
2.52−2.59 (4H, m), 2.12 (3H, s), 1.76−1.84 (2H, m), 1.21
(3H, t, J = 7.1 Hz). m/z 393 (MH+).
mixture was heated to 52
3 °C and maintained at this
temperature for 6 h. The aqueous layer was separated and
extracted with MTBE (260 L) at 52 3 °C. The combined
MTBE layers were washed with 5% aqueous sodium
bicarbonate solution (520 L) at 52 3 °C, and the aqueous
phase was re-extracted with MTBE (130 L) at 53 °C. The
combined MTBE layers were cooled slowly to 0−5 °C and held
for 12 h, and then the solid product was collected by filtration,
washed with water (260 L) followed by cold MTBE (260 L),
and dried under vacuum at 40−45 °C to provide the title
compound as a pale yellow solid, 183.7 kg (67% yield). Assay
Procedure 2. In reaction vessel 1, bis-(4-chlorobenzene)-
disulfide (12.7 g, 44.2 mmol) was suspended in ethyl acetate
(80 mL) with stirring, and the mixture was cooled to 5 °C.
Sulfuryl chloride (3.55 mL, 43.9 mmol) was added in one
portion, and the mixture was stirred at 5 °C for approximately 2
h. 9 (20.0 g, 98% w/w, 19.6 g contained weight, 83.3 mmol)
and ethyl acetate (100 mL) were charged to reaction vessel 2,
and the mixture was stirred at 20 °C. The contents of reaction
vessel 1 were added to the mixture in reaction vessel 2 over
approximately 30 min, washing in the residues with ethyl
acetate (10 mL). After stirring at briefly at 20 °C, triethylamine
(11.6 mL, 83.2 mmol) was added over a period of 12 min, and
then the mixture was held overnight. The solid byproduct was
removed by filtration; the filter cake was washed with ethyl
acetate (20 mL); and the combined filtrates were evaporated in
vacuo. The residue was dissolved in a mixture of ethyl acetate
(80 mL) and ethanol (40 mL), and then hydroxylamine
hydrochloride (5.79 g, 83.3 mmol) and tributylamine (19.9 mL,
83.6 mmol) were added. After heating at 60 °C for 4 h, the
mixture was cooled to 20 °C, and then the product was
collected by filtration, washed with ethyl acetate (2 × 40 mL),
and dried under vacuum at 40 °C to provide the title
compound as a white solid, 27.2 g (83% yield). Purity by HPLC
96.5 area %.
1
1
by H NMR 98.8% w/w. Purity by UHPLC 99.4 area %. H
NMR (500 MHz): 6.04 (1H, s), 4.79 (2H, s), 4.17 (2H, q, J =
7.1 Hz), 2.65 (2H, t, J = 6.1 Hz), 2.27 (2H, t, J = 6.3 Hz), 2.10
(3H, s), 1.94−2.03 (2H, m), 1.22 (3H, t, J = 7.1 Hz). m/z 236
(MH+).
Ethyl [(4E/Z)-4-(Hydroxyimino)-2-methyl-4,5,6,7-tetra-
hydro-1H-indol-1-yl]acetate (3). Procedure 1. In 1.0 L of
RBF, 9 (0.054 kg, 0.230 mol) was added in ethanol (0.54 L)
with stirring followed by addition of water (0.050 L),
hydroxylamine hydrochloride (0.024 kg, 0.344 mol), and
sodium acetate trihydrate (0.047 kg, 0.344 mol) in one portion.
Then the mixture was heated to reflux for approximately 3−5 h.
After cooling to 10 °C, the solid was collected by filtration on a
Buchner funnel, washed with water (0.05 L), and then dried
under vacuum at temperature of 40 °C to provide the title
compound as an off-white solid, 0.055 kg (95% yield not assay
corrected). Purity by HPLC 98.98 are a% (together for both
1
the isomers). H NMR (300 MHz): 9.99 (1H, s), 9.886 (1H,
s), 6.49 (1H, s), 5.89 (1H, s), 4.66−4.69 (4H, d), 4.11−4.19
(4H, m), 2.45−2.54 (6H, m), 2.24−2.29 (2H, m), 2.07−2.08
1
(6H, s), 1.72−1.86 (4H, m), 1.13−127 (6H, t). H NMR
Ethyl [4-Acetylamino-3-(4-chlorophenylsulfanyl)-2-
methyl-1H-indol-1-yl]acetate (6). Procedure 1. In reaction
vessel 1, acetic anhydride (23.9 kg, 234 mol) was added over a
period of 10 min to a stirred slurry of 4 (73.0 kg, 186 mol) in a
spectrum showed signals for both isomers. m/z 251 (MH+).
Ethyl [3-(4-Chlorophenylsulfanyl)-4-(hydroxyimino)-
2-methyl-4,5,6,7-tetrahydro-1H-indol-1-yl]acetate (4).
Procedure 1. In reaction vessel 1, bis-(4-chlorobenzene)-
disulfide (57.15 kg, 199 mol) was suspended in ethyl acetate
mixture of xylene (64.0 kg) and acetic acid (77.2 kg) at 22
3
°C. Meanwhile, a mixture of sodium iodide (1.39 kg, 9.3 mol),
xylene (64.0 kg), acetic acid (77.5 kg), and acetic anhydride
(23.6 kg, 231 mol) was heated to 100 3 °C in reaction vessel
2 with stirring. After stirring the contents of reaction vessel 1
for 2.5 h, this was added to the solution in reaction vessel 2
over 50 min, maintaining the temperature at between 97 and
103 °C. Reaction vessel 1 was rinsed with a mixture of xylene
(16.2 kg) and acetic acid (19.2 kg), which was added to
reaction vessel 2. The reaction mixture was held at 100 3 °C
for a further 1.5 h then cooled to 60 °C. Additional xylene (59.2
kg) was added, followed by a hot (60 °C) solution of sodium
chloride (13.7 kg) in water (106.8 kg), maintaining the
temperature at 60 °C. After stirring for 15 min, the aqueous
layer was separated and discarded. A hot (60 °C) solution of
sodium thiosulfate (2.94 kg, 18.6 mol) in water (37.2 kg) was
added, and after mixing for 20 min, the aqueous layer was
separated and discarded. Heptane (143.2 kg) was added to the
organic layer, whilst maintaining the temperature at 60 3 °C,
causing crystallisation of the product. The resulting slurry was
cooled to 20 °C over 1 h, and then the product was collected
by centrifugation, washed with ethanol (96.1 kg), and dried
(278.2 kg) with stirring, and the mixture was cooled to 0
3
°C. Sulfuryl chloride (26.9 kg, 199 mol) was added in one
portion, followed by an ethyl acetate line rinse (39.7 kg), and
then the mixture was stirred at 0 °C for approximately 2.5 h.
Meanwhile, 9 (90.0 kg, 382 mol) and ethyl acetate (351.3 kg)
were charged to reaction vessel 2, and the mixture was stirred at
20 3 °C. The contents of reaction vessel 1 were added to the
mixture in reaction vessel 2 over approximately 30 min,
washing the residues with ethyl acetate (84.5 kg). After holding
overnight at 20 °C, a solution of sodium carbonate (anhydrous,
57.7 kg, 544 mol) in water (487.3 kg) was added over a period
of 30 min (gas evolution), and then stirring was continued for
45 min. The layers were allowed to separate, and the lower
aqueous phase was discarded. A solution of sodium chloride
(31.75 kg) in water (424.5 kg) was added to the organic phase.
The mixture was stirred for 40 min, and then the layers were
separated, discarding the lower aqueous phase. The organic
layer was transferred to a clean reaction vessel, followed by an
ethyl acetate line rinse (39.0 kg), and then concentrated by
distillation at atmospheric pressure until 540 kg of distillate had
1751
dx.doi.org/10.1021/op300173z | Org. Process Res. Dev. 2012, 16, 1746−1753

Organic Process Research & Development
Article
under vacuum at a maximum jacket temperature of 40 °C to
afford the crude product 6 as an off-white solid, weight 59.0 kg
(76% yield). Assay by UHPLC 94.4% w/w. Purity UHPLC
96.9 area %.
Recrystallisation of 6. Crude 6 (87.3 kg, 209 mol) was
dissolved in a mixture of acetonitrile (622 kg) and water (396.1
kg) by heating to 80 3 °C with stirring. After holding for 20
ASSOCIATED CONTENT
* Supporting Information
HPLC and UHPLC methods for AZD1981 and intermediates.
This material is available free of charge via the Internet at
http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
parhalad.sharma@astrazeneca.com.
■
min, the solution was cooled to 15
3 °C, and then the
resulting solid product was collected by centrifugation, washed
with ethanol (116.3 kg), and dried under vacuum at a
maximum jacket temperature of 40 °C to afford purified 6 as
an off-white solid, weight 80.2 kg (92% yield). Assay by
UHPLC 99.2% w/w. Purity by UHPLC 99.4 area %. 1H NMR
(400 MHz): 9.51 (1H, br s), 7.46 (1H, d, J = 7.6 Hz), 7.26−
7.35 (3H, m), 7.11 (1H, t, J = 8.0 Hz), 6.97 (2H, d, J = 8.5 Hz),
5.24 (2H, s), 4.18 (2H, q, J = 7.1 Hz), 2.39 (3H, s), 1.86 (3H,
s), 1.21 (3H, t, J = 7.1 Hz). m/z 417/419 (MH+).
Procedure 2. A mixture of sodium iodide (3.82 g, 25.5
mmol), xylene (400 mL), acetic acid (50 mL), and acetic
anhydride (67.4 mL, 713 mmol) was heated to 100 °C in vessel
1 with stirring. In vessel 2, acetic anhydride (67.4 mL, 713
mmol) was added to a stirred slurry of 4 (200 g, 509 mmol) in
xylene (450 mL) and acetic acid (50 mL) at rt. After stirring for
45 min, this mixture was added to the contents of reaction
vessel 1 over 140 min, maintaining the temperature at 100 °C.
The reaction was held at this temperature for a further 2.5 h
and then cooled to 20 °C and left overnight. The suspension
was reheated to 60 °C, and then sodium thiosulfate (8.05 g,
50.9 mmol) and water (100 mL) were added. After mixing and
allowing the layers to separate, the lower aqueous layer was
discarded, and then the organic layer was concentrated by
distillation under vacuum, removing around 370 mL of
distillate. The temperature was adjusted to 95 °C and heptane
(600 mL) added. The resulting suspension was cooled to 20 °C
over 1 h, and then the solid product was collected by filtration,
washed with ethanol (400 mL), and dried under vacuum at 40
°C to afford 6 as a pale yellow solid, weight 171.8 g (81%
yield). Purity by UHPLC 98.8 area %.
Present Address
⊥Department of Chemical and Biological Sciences, University
of Huddersfield, Queensgate, Huddersfield, HD1 3DH, U.K.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Bill Moss and Sudhir Nambiar for their support and
Phil Cornwall for providing valuable input during initial
development work.
ABBREVIATIONS
■
TBAI (tetrabutyl ammonium iodide); Ac2O (acetic anhydride);
MIBK (methyl isobutyl ketone); MTBE (methyl-t-butyl ether);
TCCA (trichlorocyanuric acid); NH2OH·HCl (hydroxyl amine
hydrochloride); EtOH (ethanol); NaBr (sodium bromide); NaI
(sodium iodide); TFAA (trifluoroacetic acid); ZnCl2 (zinc
chloride).
ADDITIONAL NOTES
We feel the intermediate 8, being carbocyclic, is hydrophobic
■
a
in nature. The compound exists as salt in basic pH probably
because of the hydrophobic nature, and even in salt form, the
solubility of this is not there in both organic and aqueous
medium. As these facts are not confirmed, we have not added
any comments on the reason for oiling. Further telescoping
resulted in 3 without any problem. Therefore, our attention did
not go into addressing this issue.
b
The reaction aimed to have the cyclohexane-1,3-dione
consumption up to around 97%, but the conversion of 7 to 8
is around 85% by LC area. Some amount of 9a is also observed
at this stage along with de-esterified impurity which was
removed during the workup. Hence, the overall yield is around
62−67% for this two-stage process, starting from cyclohexane-
1,3-dione.
[4-Acetylamino-3-(4-chlorophenylsulfanyl)-2-methyl-
1H-indol-1-yl]acetic Acid (1). A mixture of 6 (99.9 kg, 240
mol), ethanol (393 kg), water (450 kg), and aqueous NaOH
(10 M, 72.3 kg) was heated to 62
3 °C and held at this
temperature for 30 min. The resulting solution was cooled to
20 3 °C and then filtered to remove particulate matter, and
the filter was rinsed with water (50.9 kg). MIBK (403 kg) was
added to the combined filtrates, and the mixture was heated to
60 5 °C. A mixture of aqueous hydrochloric acid (10 M, 65.0
kg) in water (296 kg) was added to the hot solution over a
period of around 45 min, maintaining the reaction temperature
within the specified range. The resulting slurry was cooled to 15
3 °C over approximately 60 min and held at this temperature
overnight. The solid product was collected by centrifugation,
washed with water (386 kg) followed by ethanol (249 kg), and
then dried under vacuum at a maximum jacket temperature of
60 °C to give the title compound as a white solid: 79.1 kg
(94%). Assay by HPLC 99.8% w/w. Purity by HPLC 99.5 area
c
Hydroxylamine hydrochloride is used for the reaction instead
of the free base to minimise the thermal hazard. Further, the
reaction calorimetric study of the oximation reaction had
shown no exotherm, indicating the safety of the process.
REFERENCES
■
(1) Bonnert, R.; Rasul, R. Novel substituted 3-sulfur indoles. WO
2004106302 A1, 2004.
(2) Moskalev, N.; Makosza, M. Tetrahedron Lett. 1999, 40 (29),
5395−5398. Makosza, M. Pure Appl. Chem. 1997, 45 (3), 559.
(3) Conley, R. T.; Ghosh, S. In Mechanisms of Molecular Migrations;
Thyagarajan, B. S., Ed.; Interscience: New York, 1971; Vol. 4, p 251.
(4) Ainge, D.; Butters, M.; Merifield, E.; Ramakrishnan, R.; Rayapati,
R.; Sharma, P.; Thomson, C. Intermeditaes and processes for the
preparation of 4-(acetylamino)-3-[(4-chlorophenyl)thio]-2-methyl-
1H-indole-1-acetic acid. WO2011004182A1, 2011.
1
%. H NMR (400 MHz): 13.20 (1H, br s), 9.51 (1H, s), 7.46
(1H, d, J = 7.5 Hz), 7.35−7.26 (3H, m), 7.11 (1H, t, J = 8.0
Hz), 6.98 (2H, d, J = 8.6 Hz), 5.12 (2H, s), 2.39 (3H, s), 1.85
(3H, s). m/z 389/391 (MH+).
(5) Block, M. H.; et al. J. Med. Chem. 2002, 45, 3509.
(6) Tamura, Y.; Nishikawa, O.; Shimizu, T.; Akita, M.; Kita, Y. Chem.
Ind. (London, U. K.) 1975, 922.
1752
dx.doi.org/10.1021/op300173z | Org. Process Res. Dev. 2012, 16, 1746−1753

Organic Process Research & Development
Article
(7) Koh, Y. K.; Bang, K.-H.; Kim, H.-S. J. Heterocycl. Chem. 2001, 38,
89.
(8) Chacon-Garcia, L.; Martinez, R. Eur. J. Med. Chem. 2002, 37, 261.
Barraja, P.; Diana, P.; Lauria, A.; Montalbano, A.; Almerico, A. M.;
Dattolo, G.; Cirrincilone, G.; Viola, G.; Dall’Acqua, F. Bioorg. Med.
Chem. Lett. 2003, 13, 2809.
(9) Janin, Y. L.; Bisagni, E. Tetrahedron 1993, 49, 10305. Carabateas,
P. M.; et al. J. Heterocycl. Chem. 1984, 21, 1849. El-Sheikh, M. I.; Cook,
J. M. J. Org. Chem. 1980, 45, 2585. Strasser, M.; Cooper, P.; Dewald,
B.; Payne, T. Helv. Chim. Acta 1988, 71, 1156.
(10) Newman, M. S.; Hung, W. M. J. Org. Chem. 1973, 38, 4073.
(11) Ishikawa, T.; Arai, M.; Nishiuchi, H.; Ishiguro, H.; Saito, S.
Chem. Lett. 2008, 37, 1066.
(12) Weidner, J. J.; Weintraub, P. M.; Schnettler, R. A.; Peet, N. P.
Tetrahedron 1997, 53, 6303. Smith, W. B.; Sant, M. P. Org. Prep.
Proced. Int. 1990, 22, 501.
(13) Sasaki, K.; Castle, R. N. J. Heterocycl. Chem. 1992, 29, 1613.
(14) Tamura, Y.; Yoshimoto, Y.; Kunimoto, K.; Tada, S-i.;
Matsumura, S.; Murayama, M.; Shibata, Y.; Enomoto, H. J. Med.
Chem. 1981, 24, 43. Beringer, F. M.; Ugelow, I. J. Am. Chem. Soc. 1953,
75, 2635.
(15) Tamura, Y.; Yoshimoto, Y.; Sakai, K.; Kita, Y. Synthesis 1980,
483.
(16) Tamura, Y.; Yoshimoto, Y.; Sakai, K.; Haruta, J.-i.; Kita, Y.
Synthesis 1980, 887.
(17) Quiclet-Sire, B.; Tolle, N.; Zafar, S. N.; Zard, S. Z. Org. Lett.
2011, 13, 3266.
(18) Tamura, Y.; Fujita, M.; Chen, L. C.; Ueno, K.; Kita, Y. J.
Heterocycl. Chem. 1982, 19, 289.
(19) Hall, C. D.; Tweedy, B. R.; Lowther, N. Phosphorus, Sulfur
Silicon Relat. Elem. 1997, 123, 341. Meinhold, H.; Muehlstaedt, M.;
Muslih, R. M. J. Prakt. Chem. 1989, 331, 136.
(20) Theobald, P. G.; Okamura, W. H. J. Org. Chem. 1990, 55, 741.
Baldwin, J. E.; Norris, R. K. J. Org. Chem. 1981, 46, 697.
1753
dx.doi.org/10.1021/op300173z | Org. Process Res. Dev. 2012, 16, 1746−1753